Why a former ARM executive is leading a UK AI startup

UK AI startups are enjoying their moment in the sun. As I reported last week, AI startups attracted $2.1 billion in VC cash in the first half of 2024, and according to a Dealroom/HSBC Innovation Banking report, the stars are aligning for record investment for the year.

But despite the influx of VC cash, UK AI innovators face a daunting challenge in developing and commercialising their technologies and solutions at a time when startups in the US, China and elsewhere are also courting customers. And of course, at the top of the tech tree, companies like IBM, Google, Microsoft and OpenAI are spending billions of dollars to stay ahead of the game.

So how does a relatively small start-up company make the journey from lab to market?

Earlier this week, I spoke with Noel Hurley, an IT veteran who spent 20 years working for processor designer ARM. During his time at the company, he held vice president positions in several divisions, including CPU and Incubation. In January of this year, he was named CEO of Literal Labs, a spinout from the University of Newcastle in northeastern England that is just beginning its commercialization journey. Given his experience working for that rarest of beasts—a British technology company that has found real success in the global marketplace—I was curious to hear his thoughts on the possibilities for Literal Labs and other research-driven AI candidates.

Finding a market niche

Let’s start with Literal Labs. Hurley believes that Literal Labs has a compelling proposition for potential customers. Instead of using the neural network technology that underpins much of today’s advanced AI, the company built a concept known as a Tsetlin machine. For those of us unfamiliar with computer science theories, all we really need to know is that it’s a technology based on a concept known as propositional logic. According to Literal Labs, it can produce tools that are up to 1,000 times more energy-efficient than their neural network counterparts.

Does it matter? Well, as Forbes.com recently reported, AI demands could be pushing the world toward an energy crisis. This is usually seen as a huge increase in consumption putting pressure on networks, but perhaps it also creates problems for companies that could benefit from embedding AI in their systems.

Hurley cites the example of manufacturing lines where AI could be used to optimize processes and increase efficiency—the holy grail of the AI ​​revolution—but the costs could be prohibitive.

So Literal Labs is focusing commercially on edge computing—a type of processing that happens not in remote server farms but closer to the action. That’s where Hurley sees a foothold.

“There’s a sector of AI called edge AI, which is applying AI to everyday devices. That could be either a consumer perspective or an industrial perspective,” Hurley says.

Beyond manufacturing lines, the company sees opportunities in self-driving cars, robots, and consumer devices. Additionally, Hurley says the Tsetlin machine concept is more transparent in terms of tracking relationships between inputs and outputs. As such, it lends itself to demand for so-called “explainable AI.” For example, say a tool uses a wide range of data to approve or reject mortgage applications. In regulated industries, there is growing demand for technologies that can explain the processes that go into making those decisions.

So there are niche markets to address—perhaps essential for a start-up company—but if solutions based on the Tsetlin machine’s concepts are a panacea, why isn’t the technology more widely used? Hurley acknowledges that there is a trade-off. On the one hand, there is speed and efficiency. On the other, processing is currently less accurate. The emphasis is now on designing for higher accuracy while finding applications where current accuracy levels are “good enough.”

Building self-confidence

In the meantime, the daunting prospect of finding customers and use cases looms. As Hurley acknowledges, it will take a lot of conversation to establish relationships and, first, build trust in the technology.

“Building a community around this technology will be important, as will building partnerships around this technology. Partnerships with neighbors in the value chain will also be important. As you know, there will be a certain amount of evangelism that will make engineers feel comfortable and enthusiastic about this technology and want to build and develop models using this approach,” he says.

As such, it is a long-term effort and the same is likely to apply to many of the AI ​​companies currently being launched at UK universities.

Holes in the road

And the road ahead is uncertain. The surge in AI investment could turn out to be a bubble that eventually bursts, making it harder for new AI companies to get traction and raise capital. Then there’s the nature of the investment. When I spoke with Yoram Winjgaarde, founder of intelligence provider Dealroom, earlier this month about AI investment in Europe, he noted, it’s not just VCs that are interested in AI. “Big tech companies are a huge part of AI investment,” he said.

That’s all well and good, but the investment in British and European AI by big tech companies could undermine the ability of companies on this side of the pond to develop and grow on their own. Hurley spoke about this in a recent interview with Fortune, in which he warned that British AI risks becoming a mere sidekick to American giants.

So what are the chances that a small research company will find a real application?

Hurley goes back to his experience at ARM, where the same groundwork had to be done. “When I was at ARM in Phase II, you could go into any semiconductor company and have an audience right away. That wasn’t the case in the early days,” he says.

But what ARM did have early on was an ambition to win global business. Hurley cites the leadership mantra of the company’s first CEO, Robin Saxby. “And he was very emphatic about the fact that we are a global company. We will put our time and effort into our customers and we will focus on the world. It’s not about being the best in the UK, it’s about being the best in the world,” Hurley says.

Finding the gaps

But given the cloud of existing big tech companies, is there room in the market for companies emerging from UK universities? Hurley says the opportunity – at least initially – lies in identifying markets and use cases that don’t interest the tech giants.

“Now, a lot of these large organizations won’t look at a whole bunch of niches at this stage, simply because they’re chasing the big contracts and the ones that are going to give them the most return for the least amount of effort,” he says. “There are often gaps that they’re not looking at. It’s about looking at the technology, what are the benefits of the technology? And honestly, going out and talking to customers who we think will benefit from our technology.”

Like many AI startups, Literal Labs is in its early days, and like any early-stage business, no one can predict whether its offering will capture the imagination of customers. But Hurley says there’s still room for emerging companies.